Study setting and design

An institution-based cross-sectional study was conducted from July 04 up to 18, 2021in the prisons of Arba Minch and Jinka town, Ethiopia. Arba Minch Town, the capital of the Gamo Zone, is 275 km from the regional town of Hawassa and 450 km from Addis Ababa. Jinka Town, the South Omo zone’s center, is 520 km from the regional town of Hawassa and 750 km from Addis Ababa. The Arba Minch Town Correctional Institution was founded in 1955E.C. and it now houses 1156 inmates. The Jinka Town Correctional Institution was founded in 1967 E.C. and now houses 760 inmates.

Sample size determination and sampling procedure

The study population consisted of all prisoners housed at Arba Minch and Jinka Town prisons. The study excluded prisoners who were openly imprisoned, seriously ill during data collection, and those who were awaiting trial. A total of 650 sample size was computed by the single population proportion formula by assuming a 4% margin of error, 95% confidence interval, and 56.4% proportion of depression among prisoners in Hawassa prison, Southern Ethiopia [12] and 10% non-response rate. All inmates in Arba Minch and Jinka town were included in this study. The proportional to size allocation was applied for each prison after the total numbers of prisoners were identified in each prison. Also, the female and the male prisoners were proportionally allocated according to their size. Study participants were selected by simple random sampling from the sampling frame obtained from registration books at the prison.

Study variablesDependent variable

Anxiety.

Independent variables

Socio-demographic related variables: Age, sex, marital status, residence, educational status, occupational status, and having children.

Behavioral-related variables: Current substance use such as alcohol use, cigarette smoking, khat chewing and using smokeless tobacco, and physical exercise.

Stressful life event and clinical disease-related variables: Stressful life event, history of chronic illness, and family history of mental illness.

Prison situation-related variables: Social support, history of previous incarceration, lack of job in prison, duration of imprisonment, and duration of punishment.

Operational definitionsAnxiety

For measuring anxiety, the Generalized Anxiety Disorder 7-item (GAD-7) scale was used with high sensitivity of 89% and a specificity of 82%. An individual with a score of 10 or more was considered to have anxiety which was the obtained cutoff value [9, 13].

Level of social support

For this study social support was measured using the Oslo 3-item social support scale. The sum score ranges from 3 to 14, with high values representing strong levels and low values representing poor levels of social support. The OSSS-3 sum score can be operationalized into three broad categories of social support with a score of 3–8 indicating poor support, 9–11 moderate support, and 12–14 strong support [14, 15].

Presence of stressful life events

Refers to Individuals who had at least one or more stressful life events (close family member died, divorce, serious illness or injury in the family member, etc.…..) in the last four weeks. The LTE contains 12 categories of significant life events, for example relating to the death of close persons, loss of relationships, imprisonment, and being the victim of theft. These 12 categories accounted for two-thirds of all events collected in the original development of the tool. Thus the experience of life events was grouped into three categories (none; 1–2 life events and 3 and above) [16].

Chronic illness

Having one disease like diabetes, hypertension, eye problem, heart disease, HIV, or kidney disease. Reported from the patient that is confirmed by the health professional.

Data collection tool and quality control

Pretested structured interviewer-administered questionnaire which had five parts was used for data collection. These were socio-demographic, behavioral factors, clinical disease-related factors, prison situation-related factors, and GAD-7. Each question on GAD-7 measures a problem that the prisoners were bothered by in the last 15 days. Anxiety symptom was rated on a scale ranging from zero (not at all) to three (nearly every day). Anxiety total scores were computed for every one of the participants by adding scores of all the seven items of the scale. A participant was considered to be in a state of depression if he/she scored ten and above [7, 13]. The stressful life event will be assessed by using an adapted version of the List of threatening experiences (LTE) questionnaire. It is a twelve items instrument measuring common life events that tend to be threatening. This instrument has been adapted and used previously in the Ethiopian setting [16, 17]. The behavioral-related questionnaire was adapted from the WHO Stepwise approach. Behavioral related focused on substance use and history of leisure-time physical exercise. Finally, part of the questionnaire asked about prison situation-related questions like (previous incarceration, duration of imprisonment, duration of punishment, and income generating jobs in prison) and social support. The tool was first developed in the English language and translated to the Amharic language then back to the English language to check the consistency. Ten (10) data collectors and six (6) supervisors (five data collectors and three supervisors for each correctional institution) who were health professionals who had direct experience and the ability to speak the Amharic language were recruited for data collection. Data quality was ensured during instrument development, collection, coding, and analysis. The supervisors and data collectors were trained about the purpose of the study and how to supervise and collect questionnaires respectively. The tool was pre-tested on 5% of the study participants in a correctional institution of Sawla correctional institution (which is not included in this study), before the actual data collection. During data collection, the questionnaire was checked for its completeness daily by supervisors and then by investigators. Intensive supervision was carried out by the assigned trained supervisors and research team members throughout the data collection period. This helped to identify problems that were addressed in the questionnaires.

Data processing and analysis

Data was collected using ODK and converted to excel then exported to SPSS version 25 for analysis. The frequency distribution of socio-demographic factors was studied after data cleaning and modification. Then, for each independent variable and outcome of interest, a binary logistic regression analysis was used to find independent factors linked to the outcome. After the binary logistic regression analysis was completed, variables with a p-value of less than 0.25 were chosen for the multiple logistic regression analysis. Context and previous studies were also taken into account while selecting a variable for multivariable analysis. The p-value associated with each parameter was evaluated with a cut-off point of less than 0.05 to determine whether or not a variable is significant. The crude and adjusted odds ratios, as well as their 95% confidence intervals, were calculated and interpreted as needed. Hosmer and Lemeshow tests were used to verify the final model’s goodness of fit.